Co-Principal Investigators: Charles Shannon (St. Louis Public Schools); Sherri Lauver (Synergy Enterprises, Inc.)

Partner Institutions: Concentric Research & Evaluation; St. Louis Public Schools; and HomeWorks! (a not-for-profit organization that delivers an intervention by the same name)

Purpose: The increasing popularity of teacher home visit programs suggests that many educators believe home visits can make a difference in improving school attendance and student achievement outcomes. This study will evaluate the effects of HomeWorks! (HW!), a teacher home visit program that has operated in the greater St. Louis area and surrounding communities since 2007, using a cluster randomized control trial (RCT) design.

Project Activities: The study will randomly assign elementary school classrooms to receive the HW! intervention or to a comparison condition in which HW! implementation is delayed. Researchers will analyze the effects of HW! on student reading achievement, days absent from school, and number of disciplinary actions.

Products: The research team will prepare separate reports for researcher and practitioner audiences. The research team, in partnership with the school district, will disseminate preliminary results to the St. Louis Public School administration. Papers will be presented at research and practitioner conferences, and will be submitted for publication in peer-reviewed journals.

Structured Abstract

Setting: The data for the study come from students in 10 elementary schools in the St. Louis Public School System, in St. Louis, Missouri. All participating schools receive federal Title I funding.

Population/Sample: The analytic sample will include data from an estimated 700 youth within 50 first- and second-grade classrooms in 10 participating elementary schools. Nearly all students will be eligible for free or reduced-price lunch and will be predominantly African American, as reflected in the demographic composition of all eligible elementary schools.

Intervention: The HomeWorks! (HW!) program is a teacher home visiting intervention aimed at bridging the gap between school and home. The HW! program model includes teacher participation in two training sessions, up to two home visits per student per academic year (two teachers attend each home visit), and attendance at two annual Family Dinners at the school. Two teachers per school act as site coordinators to help support teachers who participate in the program.

Research Design and Methods: The study is a cluster randomized controlled trial in which classrooms are randomly assigned to the HW! condition or to the comparison condition. The St. Louis school system will provide to the research team baseline and one-year follow-up administrative school records data for all students enrolled in study classrooms at the time of random assignment.

Control Condition: Teachers in the control classrooms will delay implementation of HW! for one year. In-class instruction is not impacted by the intervention and is therefore assumed to be business-as-usual in both treatment and control classrooms.

Key Measures: The primary outcome variables are reading achievement, days absent from school, and the number of disciplinary actions. Absences and disciplinary actions will be included in the administrative data provided to the research team, along with the reading achievement scores. The reading scores are from the standardized STAR assessment, which is given at all elementary schools in the district. STAR is a well-established, computer-adaptive literacy and numeracy assessment that measures concepts such as understanding of print, phonological awareness, phonics and word recognition, fluency, and vocabulary acquisition and use.

Data Analytic Strategy: The research team will estimate program effects for all students randomly assigned to the treatment and comparison groups (intent-to-treat analysis). Researchers will statistically control for pre-intervention measures of the outcomes as well as demographic characteristics; adjust standard errors for classroom-level clustering; and include school indicators as fixed effects. The researchers will also use multilevel modeling to conduct a treatment-on-the-treated analysis.